ORIGINAL RESEARCH published: 29 March 2018 doi: 10.3389/fmars.2018.00106

Predator in the Pool? A Quantitative Evaluation of Non-indexed Journals in Aquaculture Research

Jeff C. Clements 1*, Rémi M. Daigle 2 and Halley E. Froehlich 3

1 Aquaculture and Coastal Ecosystems Section, Fisheries and Oceans Canada, Gulf Fisheries Centre, Moncton, NB, Canada, 2 Département de Biologie, Laval University, Québec, QC, Canada, 3 National Center for Ecological Analysis and Synthesis, University of California, Santa Barbara, Santa Barbara, CA, United States

Predatory open access (OA) journals can be defined as non-indexed journals that exploit the gold OA model for profit, often spamming academics with questionable e-mails promising rapid OA publication for a fee. In aquaculture—a rapidly growing and highly scrutinized field—the issue of such journals remains undocumented. We employed a quantitative approach to determine whether attributes of scientific quality and rigor differed between OA aquaculture journals not indexed in reputable databases

Edited by: and well-established, indexed journals. Using a Google search, we identified several António V. Sykes, non-indexed OA journals, gathered data on attributes of these journals and articles Centro de Ciências do Mar (CCMAR), Portugal therein, and compared these data to well-established aquaculture journals indexed Reviewed by: in quality-controlled bibliometric databases. We then used these data to determine if Pedro Morais, non-indexed journals were likely predatory OA journals and if they pose a potential threat University of California, Berkeley, to aquaculture research. On average, non-indexed OA journals published significantly United States Enrique Orduna-Malea, fewer papers per year, had cheaper fees, and were more recently established than Universitat Politècnica de València, indexed journals. Articles in non-indexed journals were, on average, shorter, had fewer Spain authors and references, and spent significantly less time in than their *Correspondence: Jeff C. Clements indexed counterparts; the proportion of articles employing rigorous statistical analyses [email protected]; was also lower for non-indexed journals. Additionally, articles in non-indexed journals [email protected] were more likely to be published by scientists from developing nations. Worryingly,

Specialty section: non-indexed journals were more likely to be found using a Google search, and their This article was submitted to articles superficially resembled those in indexed journals. These results suggest that Marine Fisheries, Aquaculture and the non-indexed aquaculture journals identified herein are likely predatory OA journals Living Resources, a section of the journal and pose a threat to aquaculture research and the public education and perception Frontiers in Marine Science of aquaculture. Several points of reference from this study, in combination, may help Received: 26 July 2017 scientists and the public more easily identify these possibly predatory journals, as Accepted: 14 March 2018 these journals were typically established after 2010, publishing <20 papers per year, Published: 29 March 2018 < < Citation: had fees $1,000, and published articles 80 days after submission. Subsequently Clements JC, Daigle RM and checking reputable and quality-controlled databases such as the Directory of Open Froehlich HE (2018) Predator in the Access Journals, , , and Thompson Reuters can aid in confirming Pool? A Quantitative Evaluation of Non-indexed Open Access Journals in the legitimacy of non-indexed OA journals and can facilitate avoidance of predatory OA Aquaculture Research. aquaculture journals. Front. Mar. Sci. 5:106. doi: 10.3389/fmars.2018.00106 Keywords: ethics, journal selection, open access, peer review, scientific publishing, science communication

Frontiers in Marine Science | www.frontiersin.org 1 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

INTRODUCTION use questionable peer review practices (Bohannon, 2013), and publish low quality science and/or non-scientific information In addition to dramatic environmental change, the (Bohannon, 2013; “shoddy science indeed, they publish.”— Anthropocene Epoch is also characterized by rapid and drastic Dockrill, 2017). Thus, predatory OA journals can disseminate human societal changes (Ellis et al., 2016). One such societal potentially dubious results that have not received adequate peer shift is the ways that humans communicate with one another, as review, fundamentally threatening principled scientific integrity advances in technology have created a hyper-connected world in (Beall, 2016; Vinny et al., 2016). which considerable amounts of information are available at the Researchers may be naïve to predatory OA publishers (which, push of a button. As such, the ways that humans view and interact for some fields, can be an alarming percentage of researchers; with the natural world and one another in this time of great Christopher and Young, 2015) or can be pressured to publish change can be impacted by the information that they receive and may thus use predatory OA journals as an outlet for and the ways in which it is perceived (Castree, 2017; Holmes disseminating their research quickly (Van Nuland and Rogers, et al., 2017). Unfortunately, however, this hyper-connectivity 2016). For naïve but well-intended researchers, publishing in has resulted in a platform for the dissemination of propagandist predatory OA journals might hinder career progression if hiring information and the onset of a “post-truth era” (e.g., Keyes, 2004; committees penalize researchers for publishing in such journals. d’Ancona, 2017). Thus, now, more than ever, we rely on the Alternatively, unethical researchers can also exploit predatory adequate dissemination and propagation of objectively truthful OA journals to increase their publication numbers by avoiding information on a global scale. Given that science is arguably the rigorous peer review (Shen and Björk, 2015; Berger, 2017), best method to obtain objective truth, the scientific community and such practices can make unethical researchers appealing must endeavor to make scientific information accessible and to hiring committees by increasing their publication numbers. understandable to mass audiences outside of the scientific Furthermore, predatory OA journals can provide an outlet for community if truth and logic are to prevail in the Anthropocene. the publication of research from nations that are often excluded Open access publishing—making academic research freely (intentionally or not) from Western publishing outlets. Indeed available to everyone—has emerged as one tool to make science it is reported that the majority of authors publishing in these more accessible to all by providing universal free access to journals are researchers in developing nations who could better scientific articles. The open access approach can have many use their funds to stimulate research activities or publish in non- benefits. For example, the citation counts of open access articles predatory open access journals (Shen and Björk, 2015; Xia et al., are often higher than those of closed-access articles (Antelman, 2015; Seethapathy et al., 2016; Balehegn, 2017; Gasparyan et al., 2004; Hajjem et al., 2005; Eysenbach, 2006; Craig et al., 2007; 2017). Consequently, predatory publishing practices threaten the Gargouri et al., 2010; Clements, 2017). Open access articles are institution of science, undermine legitimate open access journals, also reported to attract more media attention and increased and hinder progression of the utilitarian concept of open science author exposure (Adie, 2014; Wang et al., 2015; McKiernan on a global scale. et al., 2016). Most importantly, the general premise of open For areas of research that are highly applicable to the access publishing allows scholarly works to be legally accessed public, it is important to elucidate ways of distinguishing by anyone with internet access (although illegal means of predatory OA journals from legitimate journals. One such area obtaining scholarly works blocked behind subscription paywalls of research that has yet to receive attention regarding predatory are available, e.g., SciHub; McNutt, 2016), making scientific OA journals is the field of aquaculture. Since the onset of the information available to those who have historically been Anthropocene, aquaculture (i.e., aquatic farming) activities have restricted (e.g., researchers in developing nations that cannot become characteristic of many coastal regions around the world. afford journal subscriptions, the general public; Björk et al., 2014; Aquaculture is one of the fastest growing food sectors in the Swan et al., 2015). world, providing a source of food for millions of people in both Although there are many benefits of OA publishing, the developed and developing nations (FAO, 2016). Aquaculture approach has been linked to the emergence of so-called activities are also a major component of the global economy, with “predatory OA journals” (Beall, 2012, 2013), which have rapidly an estimated value of more than USD$160 billion (FAO, 2016). increased in recent years (Shen and Björk, 2015). Predatory Furthermore, while capture fisheries have ceased growth (Pauly OA journals are typically not indexed in quality-controlled and Zeller, 2016) the aquaculture sector continues to increase at databases [e.g., Directory of Open Access Journals (DOAJ), Web an astonishing rate, now accounting for half of all seafood and of Science, Scopus, Thompson Reuters; however it is important highlighting its importance for future economic growth and food to note that not all non-indexed OA journals are predatory] security (FAO, 2016). Although aquaculture plays a key role in and exploit the gold OA model for profit. These journals also global food systems, public perception of aquaculture activities provide substandard services, accept all submitted articles, and are often negative and accompanied by many misconceptions, tend to spam academics with disingenuous e-mails and the hindering productive and informed discussion (Fisheries and promise of rapid publication in their fully open access journals Oceans Canada, 2005; Bacher, 2015; D’Anna and Murray, 2015; for a fee (Jalalian and Mahboobi, 2014; Moher and Srivastava, Froehlich et al., 2017). It is suggested that clear communication 2015); many even report bogus and/or alternative impact factors and adequate investigation into the real and perceived threats (Gutierrez et al., 2015). Such journals are often run by and of aquaculture could help shift public perception and support recruit non-scientists (e.g., fake editors; Sorokowski et al., 2017), the growth and sustainability of aquaculture activities (Froehlich

Frontiers in Marine Science | www.frontiersin.org 2 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals et al., 2017). Thus, public accessibility to rigorous, high quality three search pages, although the order of journal appearance science is critical to stimulate productive dialogue regarding changed with each new keyword string). A Google search aquaculture activities and sustainability. was used because non-indexed journals are not referenced in While predatory OA journals can provide a seemingly- reliable searchable databases (e.g., DOAJWeb of Science, Scopus, professional platform for diminutive science and ideas across Thompson Reuters), and because the public most likely have a number of disciplines, the ways in which such journals access to and use general Google searches to obtain information threaten the field of aquaculture and influence associated public (as opposed to scholarly databases including Google Scholar; perception is unknown. Furthermore, the careless and loose Purcell et al., 2012; Fox and Duggan, 2013; Richter, 2013). We attribution of the term “predatory” to many OA journals then assessed each hit to determine if the associated webpage that are not predatory has recently been heavily criticized presented itself as a scientific journal (i.e., contained apparently (Swauger, 2017). Consequently, an objective understanding of scientific research articles pertaining to aquaculture) and limited the severity of predatory OA publishing in aquaculture is our search to journals with the term “aquaculture” in the title. We needed. A more objective understanding of predatory OA discarded hits that were not scientific journals, and also discarded aquaculture journal attributes and comparing them to well- duplicate journals within and between SERPs. established aquaculture journals would not only make it easier for Due to increasing redundancy in relevant material, we only aquaculture stakeholders (scientists, fish farmers, policy makers) searched the first 10 SERPs for scientific journals (Figure 1A). to identify predatory OA journals with more confidence, but Each journal was then categorized as well-established and would provide a measure for the severity of the problem within indexed (hereafter referred to as “indexed”), non-indexed OA, this highly scrutinized field. Moreover, comparing proxies of or unassigned (i.e., journals that could not be identified as scientific rigor between predatory and non-predatory OA journal either indexed or non-indexed OA). We initially identified articles, such as methodological rigor (e.g., statistical rigor) and non-indexed OA journals by searching three quality-controlled content quality, would begin to address whether or not such databases: the Directory of Open Access Journals (DOAJ), journals really pose a threat to aquaculture research. Clarivate Analytics (, JCR), and Scopus. The overall goal of this study was to gain a more “Indexed” aquaculture journals were characterized as journals quantitative and objective understanding of both the quality of that were listed in the DOAJ, Clarivate Analytics JCR, and/or the aquaculture journals easily found by internet search engines Scopus. Indexed journals could be partially or fully OA; fully OA and the possible associated consequences, and objectively bring indexed journals were checked in the DOAJ, Clarivate Analytics to attention the issue of predatory OA publishing within this field JCR, and Scopus to ensure that they were verified as accredited of research. More specifically, we attempted to: 1. Quantitatively OA journals. Unassigned journals were characterized as journals compare characteristics of non-indexed OA aquaculture journals that were not indexed in the DOAJ, Clarivate Analytics and the articles therein to those of well-established, indexed JCR, or Scopus, and were not published under a mandatory aquaculture journals (to determine the likelihood of non-indexed OA policy (and could therefore not exploit gold OA for OA journals being predatory), 2. Determine the probability of profit). the general public encountering non-indexed OA aquaculture journals using a Google search, and 3. Use the results of the aforementioned goals to conceptualize potential risks of Data Collection predatory OA journals to aquaculture research. We hypothesized For each of the 10 SERPs assessed, we recorded the number of that non-indexed OA journals would appear more frequently journals from each of the three categories. From these data, we in a Google search, and that articles in those journals would calculated the per-SERP and overall totals of each journal type exhibit tendencies of lesser scientific quality than those in (i.e., indexed, non-indexed OA, and unassigned) found in the well-established, indexed journals based on proxies of scientific Google search. To obtain comparative data on journal attributes rigor. between indexed and non-indexed OA aquaculture journals, we analyzed a subsample of journals by choosing the first six METHODS journals from each category (ncategory = 6, Ntotal = 12) that had a global scope (i.e., were not restricted to a certain geographic Google Search and Journal area, e.g., North America); this provided a representative sample Characterization of both journal types and a balanced sampling design for We conducted a simple Google search for potential scholarly statistical analyses. Each selected journal had to have published aquaculture journals using the keyword string “journal of at least one issue in 2017 (as well as papers in previous years; aquaculture”; advertisements and sponsored links appearing on ensuring at least one issue in 2017 provided confirmation that each search-engine results page (SERP) were excluded. This the journal was still actively publishing, but year of publication keyword string was chosen based on what we deemed to be was not taken into account when sampling articles because the most likely keyword string that individuals would use when the metrics obtained from each paper would not be affected searching for scholarly journals in aquaculture. We also searched by publication year), had published a total of >20 articles (to keyword strings “journal aquaculture”and“aquaculture journal,” accommodate our chosen article sample size; see next paragraph), which provided similar proportions of non-indexed vs. indexed and contained a peer review policy on the journal or publisher journals (13 and 12 non-indexed OA journals within the first website. For non-indexed OA journals, there also had to be

Frontiers in Marine Science | www.frontiersin.org 3 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

FIGURE 1 | (A) The total number of scientific journals (n = 48) identified on each search engine results page (SERP; 10 hits page−1). (B) The cumulative number of scientific journals identified in the Google search (across the 10 SERPs) that were categorized as non-indexed OA, indexed, or unassigned. (C) The percentage of journals identified on each SERP that were categorized as non-indexed OA (black), indexed (gray), and unassigned (white).

a clear indication that authors would need to pay a fee for OA journals published poor science and could thus potentially be publication. predatory, we recorded various proxies of scientific rigor at the For each of the six subsampled journals from each category, article level (see Table 1); it should be noted, however, that these we collected a variety of journal-level and article-level metrics proxies are not necessarily direct indicators of scientific quality (Table 1; see Tables S1–S3 for raw data). At the journal level, we and rigor. Some journals did not report the time from submission recorded the year that the journal was established (i.e., age), the to acceptance nor the time from acceptance to publication and type of peer review reported (single blind or double blind), the sample sizes were thus reduced for these metrics (n = 100 for APC for open access (where applicable), and the average number predatory and 80 for non-predatory). of papers published per year (total number of papers ÷ age). Two journals (one non-indexed OA and one indexed) did not report Statistical Analysis open access APCs, reducing the sample size for this variable from Since our sampling design included both fixed and random 6 per group to 5 per group. factors, we used linear mixed models to test for the effect of To obtain article-level metrics, we sampled the 20 most recent journal type (fixed factor) on journal-level metrics (ncategory articles that were assigned to a volume and issue (irrespective = 6) and article level metrics (ncategory = 120 except for of publication year; njournal = 20, ncategory = 120, Ntotal = 240). proportional data with ncategory = 6). For journal level metrics For article-level metrics, we recorded the type of article (research, and proportional article-level metrics, “journal” was included review, or commentary), the number of authors, the nationality as a random factor to control for any random effects of of the corresponding author and all other authors, and the individual journal; likewise, article nested within journal was economic classification of an author’s country (developed, in included as a random variable for non-proportional article level transition, or developing, as defined by the UN; United Nations, metrics. AIC/log likelihood tests comparing models with and 2017). To quantitatively determine whether or not non-indexed without the random factors were used to determine whether

Frontiers in Marine Science | www.frontiersin.org 4 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

TABLE 1 | Descriptions and associated categories (where applicable) for each journal- and article-level metric collected.

Metric Description Categories

JOURNAL LEVEL Year of inception The year in which the first issue of the journal was published – Peer review type The type of peer review employed by the journal. If not explicitly stated, assumed to be single Single blind / Double blind blind. OA fee (APC) The cost to publish an article as open access in a given journal (in USD$) – Average papers per year The average # of papers per year (total # papers÷# of years established) – ARTICLE LEVEL Number of authors The total number of authors – Author nation Country of authors’ affiliations; assessed separately for corresponding authors and other – authors Author nation type The economic classification of authors’ affiliations; assessed separately for corresponding Developed/In authors and other authors transition/Developing Article length Total number of pages – Time from submission to acceptance Number of days from submission to acceptance – Time from acceptance to publication Number of days from acceptance to publication – Number of references The total number of references cited in the article – Number of figures and tables The total number of figures in the paper – Statistics reported? Whether or not the authors reported any statistical analyses in the methods and results (must Yes/No have been reported in both for “yes”); studies not reporting any quantitative data (e.g., review papers, commentaries) were excluded from this analysis Detailed statistics reported? Whether or not statistical details were reported. Must have used an appropriate test and Yes/No included at least: (i) recognition and appropriate treatment of statistical assumptions, or (ii) reporting of full statistical results (test statistics + df + p-value); studies not reporting any quantitative data (e.g., review papers, commentaries) were excluded from this analysis Downloadable PDF? Whether or not the article is available as a downloadable PDF. Yes/No

or not random effects were evident (Burnham and Anderson, journals by promising quick publication and comparatively 2002). Assumptions of normality and homoscedasticity were cheap OA fees (indicated by significantly shorter times between assessed visually using Q-Q plots and residual plots, respectively. submission and publication, and significantly cheaper APCs, in Journal-level metrics were log transformed prior to analyses to non-indexed OA journals), and 3. Exploiting naïve scientists avoid violations of homoscedasticity and all proportion data and/or providing an easily exploitable platform for unethical were arcsin square root transformed prior to analyses [Ahrens (“predatory”) authors from developing nations, where the et al., 1990; while scrutinized (Warton and Hui, 2011), arcsin pressure to publish is often intense (Butler, 2013). Each of the transformations allowed the data to meet statistical assumptions]. three negative impacts to science that were found to hold true All statistical analyses were conducted in R version 3.4.1 (R from our quantitative assessment were then translated into direct Development Core Team, 2017) using the “nlme” package effects to aquaculture research. (Pinheiro et al., 2017) with a significance threshold of α ≤ 0.05. R-code for all analyses can be found on GitHub and has been RESULTS archived on Zenodo (see Supplementary Material section). Google Search Conceptualization of Predatory OA Journal Non-indexed OA aquaculture journals were ca. three times more Impacts on Aquaculture Research numerous and common in a Google search than indexed and Using the results of the quantitative assessment, we derived unassigned journals (Figure 1). Across the 10 SERPs assessed, a conceptualization of the potential negative impacts of we identified a total of 30 non-indexed OA journals, 8 indexed predatory OA journals on aquaculture research. For impact journals, and 10 unassigned journals (Figure 1B). While not conceptualization, we first identified three potential negative strictly OA, the majority of indexed journals were considered impacts of predatory OA journals and subsequently determined hybrid journals (i.e., offered the option of making individual whether or not these results supported or refuted the existence articles OA, but not requiring it); only one indexed journal of such impacts in the field of aquaculture: 1. Lack of peer had a mandatory OA policy and granted waivers and discounts review (indicated by significantly shorter durations between to researchers from underdeveloped nations. Importantly, the submission and acceptance in non-indexed OA journals, along proportion of non-indexed OA journals was equal to (3 of 10 with significantly lower indicators of scientific and statistical SERPs) or higher than (6 of 10 SERPs) that of indexed journals rigor), 2. Enticing scientists to publish work in predatory OA on all SERPs with the exception of SERP eight (Figure 1C).

Frontiers in Marine Science | www.frontiersin.org 5 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

Journal Level Metrics Articles in non-indexed OA journals had significantly fewer We found significant differences between indexed and non- authors [indexed: 4.69 ± 0.19, non-indexed OA: 3.38 ± 0.42 indexed OA journals for all three journal-level metrics (Figure 2, authors; F(1, 10) = 11.02, p = 0.008], were significantly shorter Table 2), much of it rooted in the longevity of the journals [indexed: 12.70 ± 0.65, non-indexed OA: 6.99 ± 0.60 pages; themselves. Relative to indexed journals, non-indexed OA F(1, 10) = 17.61, p = 0.002], and were less well referenced aquaculture journals were characterized by more recent inception [indexed: 49.73 ± 2.03, non-indexed OA: 36.32 ± 4.84 < years [minimum year of inception: non-indexed OA = 2010, references; F(1, 10) = 22.42, p 0.001] than articles in indexed < indexed = 1970; F(1, 10) = 26.61, p 0.001; Table 2, Figure 2A]. journals (Figures 3A–C, Table 3). In contrast, articles from In fact, the indexed journal with the most recent inception date indexed and non-indexed OA journals had similar numbers (2009) was still older than the oldest non-indexed OA journal of figures and tables (Figure 3D, Table 3). While non-indexed (2010). The older age of indexed journals was associated with OA articles spent significantly less time under peer review than higher numbers of articles, volumes, and issues than non-indexed articles in indexed journals [indexed: 150.25 ± 6.91, non-indexed OA journals. Non-indexed OA journals were also characterized OA: 75.94 ± 17.70 days; F(1, 7) = 8.29, p = 0.024] (Figure 3E, by cheaper APCs (US$991.50 ± 215.23) when compared Table 3), the time from acceptance to online publication (i.e., to indexed journals (US$2920.60 ± 268.31) [F(1, 8) =15.26, time in production) was statistically similar between articles p = 0.005; Table 2, Figure 2B], and tended to publish fewer in indexed and non-indexed OA journals [indexed: 29.77 ± −1 −1 papers year than indexed journals [21.4 ± 5.6 papers year 2.16, non-indexed OA: 26.57 ± 10.16 authors; F(1, 7) = 0.03, in non-indexed OA journals vs. 94.1 ± 42.0 in indexed journals; p = 0.877] (Figure 3F, Table 3). Additionally, the proportion F(1, 10) = 5.74, p = 0.018; Table 2, Figure 2C]. Interestingly, while of articles reporting both basic statistics [indexed: 0.95 ± 0.03, < all indexed journals surveyed for this study reported a single non-indexed OA: 0.50 ± 0.08; F(1, 10) = 33.89, p 0.001] and blind peer review policy (i.e., authors are not made aware of detailed statistical analyses [indexed: 0.71 ± 0.09, non-indexed < the reviewers’ identity), 33% (2/6) of non-indexed OA journals OA: 0.05 ± 0.03 authors; F(1, 10) = 40.54, p 0.001] was far reported double blind peer review, while an additional 33% of higher in indexed journals than non-indexed OA journals non-indexed OA journals reported that double blind review (Figures 3G,H, Table 3). While clear differences between articles would be made available at the authors’ request (i.e., 66% of in indexed and non-indexed OA journals existed for metrics non-indexed OA journals offered double blind peer review). of scientific rigor, the incorporation of journal as a random Statistically, linear mixed effects modeling revealed significant variable significantly improved model fit for many of these differences between indexed and non-indexed OA journals for metrics (number of authors, length, number of tables, and all all six journal-level metrics, while individual journal had no effect three measures of time prior to publication; Table 3). (Table 2). Non-indexed OA journals had a statistically higher number of authors from developing countries than indexed journals. Article Level Metrics Alongside displaying comparatively weaker scientific rigor, Although non-indexed OA journal articles superficially non-indexed OA journals had significantly higher [indexed: resembled articles in indexed journals, indicators of scientific 0.51 ± 0.07, non-indexed: 0.78 ± 0.08; F(1, 10) = 5.50, rigor in non-indexed OA journals were, in general, lower. p = 0.04] proportions of corresponding and/or primary authors

FIGURE 2 | Boxplot representation of journal-level metrics for indexed (gray boxes) and non-indexed OA (white boxes) journals, including the time since the journal was incepted (A), the cost of publishing open access (A.P.C.) (B), and the average number of papers published per year (C). Asterisks denote significant differences between journal types (see Table 2). n = 6 per group.

Frontiers in Marine Science | www.frontiersin.org 6 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

TABLE 2 | Results of linear mixed effects models and AIC/log likelihood model comparisons for effects of journal type and journal on journal-level metrics.

LME AIC/log likelihood comparison

numDF denDF F-value p-value df AIC BIC logLik L. Ratio p-value

YEARS SINCE INCEPTION (Intercept) 1 10 303.35 <0.001 w/random 4 26.73 27.94 −9.37 Journaltype 1 10 26.61 <0.001 w/orandom 3 26.73 25.64 −9.37 <0.001 >0.999 APC (Intercept) 1 8 3672.22 <0.001 w/random 4 18.86 19.18 −5.43 Journaltype 1 8 15.26 0.005 w/orandom 3 16.86 17.10 −5.43 <0.001 >0.999 PAPERS PER YEAR (Intercept) 1 10 282.77 <0.001 w/random 4 34.27 35.48 −13.14 Journaltype 1 10 5.74 0.018 w/orandom 3 32.27 33.18 −13.14 <0.001 >0.999

Bolded text denotes statistically significant main effects at α ≤ 0.05. from developing countries (Figure 4A, Table 4); however, the Predatory OA Publishing as a Source of proportion of additional authors from developing countries was Misinformation in Aquaculture similar between indexed and non-indexed OA journals [indexed: Our results suggest that while they are likely of lesser quality, 0.57 ± 0.07, non-indexed OA: 0.73 ± 0.10; F(1, 10) = 2.37, non-indexed OA aquaculture journals often mimic legitimate, p = 0.155; Figure 4B, Table 4]. high-quality scientific articles (e.g. similarly named, professional website, PDF articles). Thus, it is not trivial to distinguish DISCUSSION legitimate and reputable journals from those that are not. Theoretically, trained scientists should be able to recognize Can Non-indexed OA Journals in potentially predatory OA journals based on the quality of science Aquaculture Research Be Predatory? and proceed accordingly. In reality, however, this is not always To our knowledge, this is the first study to measure attributes the case; with approximately 6,800 scientific articles published of journals in aquaculture research to provide a quantitative each day (Ware and Mabe, 2015), possible predatory submissions comparison between non-indexed OA journals and well- are bound to be overlooked. Hypothetically, if researchers established, indexed journals for scientific standards, ethical simply skim a paper for information and do not recognize that rigor and possible predatory implications. Broadly, our results the paper is from a potential predatory journal and has not suggest that non-indexed OA journals are widespread within been through rigorous peer review, misinformation can become aquaculture research and tend to be less established, contain embedded within the aquaculture literature, as well as other fewer papers, and offer OA at a lower cost. Importantly, these fields. Furthermore, hypothetically, naïve researchers unfamiliar components appear to align with reduced scientific quality with the shifting landscape of scientific publication and the (fewer references, shorter peer review) and statistical rigor. Thus, threats therein may submit and publish work in potentially our results suggest that many non-indexed OA aquaculture predatory OA journals, unbeknownst to the consequences. As journals show characteristics of predatory OA journals (as mentioned previously with developing researchers, this may described in the section Introduction), and support the idea result in high-quality work being published in a predatory OA that predatory journals often (but perhaps not always) use journal, ultimately going unrecognized or being dismissed as questionable peer-review practices and publish substandard inadequate. In addition, the reputation of individual researchers science and/or non-scientific information (Bohannon, 2013) that can be influenced, depending on whether or not repeated can lead to propagation of false information. It is important publications in potentially predatory OA journals are noticed and to note, however, that our results do not show that all non- properly critiqued. As such, future studies conducting a deeper indexed OA aquaculture journals are predatory OA journals, and evaluation of the scientific content of articles in known predatory individual journals need to be assessed using a combination of aquaculture journals than we provide herein are warranted to criteria outlined below. Furthermore, not all articles published better understand the extent of this effect. At current, however, it in non-indexed OA journals are necessarily of poor quality. is unclear just how much of the aquaculture research community Nonetheless, our results provide evidence that predatory OA (and the industry at large) is naïve to predatory OA journals. A journals may be affecting aquaculture science through a lack quantitative documentation of the susceptibility of researchers to of proper peer review (i.e., short times between submission predatory OA publishing is necessary to fully comprehend how and acceptance which could lead to increased pressure on peer much of a risk to aquaculture predatory publishers pose. reviewers and promote errors in peer review), enticing scientists As mentioned, the non-indexed OA aquaculture journals to publish work in predatory journals by promising quick highlighted in our study used less rigorous peer review (i.e., publication and comparatively cheap OA fees, and by exploiting shorter peer review times which could put pressure on reviewers (either advertently or inadvertently) scientists in developing to review papers quickly, thus promoting errors and poorer nations (Figure 5). peer review) and scientific work, they superficially resembled

Frontiers in Marine Science | www.frontiersin.org 7 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

FIGURE 3 | Boxplot representation of non-demographic article level metrics for indexed (gray boxes) and non-indexed OA (white boxes) journals, including the number of authors (A), the article length (in pages) (B), the number of references (C), the number of figures and tables (D), the time (in days) from submission to acceptance (E) and the time from acceptance to publication (F), and the proportion of articles reporting basic (G) and detailed statistical analyses (H). Asterisks denote significant differences between journal types; ns denotes non-significant differences (see Table 3). See Methods section Data collection for details on sample sizes. well-established, indexed journals. Their existence thus poses a Benkendorff, 2009), and predatory OA journals are known to potential threat to the communication of aquaculture research publish false pharmaceutical findings in other fields (Bohannon, to the public. Although most scientists are trained to critically 2013). Thus, predatory OA aquaculture journals are not only evaluate individual scientific articles, the same cannot be a potential threat to the academic side of aquaculture, but said for non-scientists, as non-scientists are typically not well also to the public perception of aquaculture activities and the versed to distinguish between high- and low-quality science. economic and ecological sustainability of the industry. It should Astonishingly, the non-indexed OA journals herein were three be noted, however, that well respected, non-predatory publishers times more abundant than well established, indexed journals can also publish faulty and/or non-scientific content [e.g., the in our Google search. This is concerning when one considers autism-vaccine link (Wakefield et al., 1998), or the recent high- that a large portion of the public obtains much of its scientific profile case of microplastic effects on perch larvae (Lönnestedt information from Google searches (Purcell et al., 2012; Fox and and Eklöv, 2016)] and the spread of misinformation within Duggan, 2013). It is important to recognize, however, the Google the field of aquaculture research is not restricted to predatory searches are personalized, and Google results will vary across OA journals. Nonetheless, our results do suggest that predatory users and over time; a more in depth study assessing Google OA aquaculture journals have a higher propensity to publish search results across a broad range of users and time periods lesser-quality science than their non-predatory counterparts and is thus necessary to confirm the findings herein. Nonetheless, can potentially contribute to the spread of misinformation in our search was adequate in identifying non-indexed OA journals aquaculture. for quantitative comparison with indexed journals. Furthermore, fish farmers and other stakeholders in the aquaculture industry rely on sound scientific information to inform optimal practices Who Publishes in Non-indexed OA in order to ensure economic and environmental optimization Aquaculture Journals? and stability. For example, the treatment of farmed fish Our results agree with those of previous studies suggesting that with tested and approved pharmaceuticals can be critical non-indexed OA journals serve a disproportionate number of in maintaining healthy and high-quality fish (Shao, 2001; authors from developing countries (Shen and Björk, 2015; Xia

Frontiers in Marine Science | www.frontiersin.org 8 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

TABLE 3 | Results of linear mixed effects models and AIC/log likelihood model comparisons for effects of journal type and journal on non-demographic article-level metrics.

LME AIC/log likelihood comparison

numDF denDF F-value p-value df AIC BIC logLik L.Ratio p-value

NUMBER OF AUTHORS (Intercept) 1 228 428.88 <0.001 w/random 4 994.85 1008.74 −493.43 Journaltype 1 10 11.02 0.008 w/orandom 3 999.13 1009.55 −496.57 6.28 0.012* ARTICLE LENGTH (Intercept) 1 228 208.01 <0.001 w/random 4 1478.86 1492.75 −735.43 Journaltype 1 10 17.61 0.002 w/orandom 3 1493.65 1504.07 –743.82 16.79 <0.001* NUMBER OF REFERENCES (Intercept) 1 228 923.23 <0.001 w/random 4 2162.91 2176.78 –1077.45 Journaltype 1 10 22.42 <0.001 w/orandom 3 2171.33 2171.33 –1077.45 <0.001 >0.999 NUMBER OF FIGURES AND TABLES (Intercept) 1 207 487.05 <0.001 w/random 4 1202.45 1215.97 –597.22 Journaltype 1 10 1.31 0.280 w/orandom 3 1201.50 1211.64 –597.75 1.05 0.305 TIME: SUBMISSION TO ACCEPTANCE (Intercept) 1 171 72.21 <0.001 w/random 4 2053.09 2065.82 −1022.55 Journaltype 1 7 8.29 0.024 w/orandom 3 2069.73 2079.28 –1031.87 18.64 <0.001* TIME: ACCEPTANCE TO PUBLICATION (Intercept) 1 171 8.43 0.004 w/random 4 1726.69 1739.49 −859.35 Journaltype 1 7 0.03 0.877 w/orandom 3 1809.04 1818.58 –901.52 84.34 <0.001* PROPORTION OF ARTICLES REPORTING BASIC STATISTICS (Intercept) 1 10 184.78 <0.001 w/random 4 11.34 12.55 –1.67 Journaltype 1 10 33.88 <0.001 w/o random 3 9.34 10.25 –1.67 <0.001 >0.999 PROPORTION OF ARTICLES REPORTING DETAILED STATISTICS (Intercept) 1 10 52.65 <0.001 w/random 4 8.65 9.86 –0.33 Journaltype 1 10 40.54 <0.001 w/o random 3 6.65 7.56 –0.33 <0.001 >0.999

Bolded text denotes statistically significant main effect of journal type from linear mixed effects model at α ≤ 0.05; asterisk denotes significant difference between models with and without random factor (where significant, the best model has the lowest AIC and highest log likelihood, and is italicized). et al., 2015; Seethapathy et al., 2016; Balehegn, 2017; Gasparyan has yet to be quantitatively assessed) and can be exploited by et al., 2017). It thus seems that many aquaculture researchers unethical researchers for personal gain (Shen and Björk, 2015), from these regions are spending a fair proportion of their aquaculture research from such nations may be perceived by the limited financial resources on publishing in likely predatory OA academic community as low quality, even though it may not be. journals, when such resources can be used in more effective As such, the publication of aquaculture research in non-indexed ways (e.g., increasing research infrastructure, publishing in high OA journals may hinder the research reputation of nations quality open access journals, hiring staff). This may be due to a that publish in such journals often. To combat this, individual number of factors such as preferential targeting of researchers researchers should always use the “think, check, submit” method in developing-nations by predatory OA publishers, a shorter (see next section) when submitting papers to OA journals. In cultural tradition of participation in scientific publication, an addition, indexed journals and other relevant institutions should intense pressure to publish resulting in unethical publishing explore ways to remove publication barriers for researchers in practices (i.e., predatory authors), and/or economic limitations developing nations and increase incentives for such researchers competing against high APCs in non-predatory OA journals to publish their work in reputable journals. (Harris, 2004; Butler, 2013; Shen and Björk, 2015). It is important to note, however, that particularly the latter factor is not at Minimizing the Impact of Potential all limited to developing nations. It is also likely that good, Predatory OA Publishing in Aquaculture high-quality research sometimes gets published in questionable Alongside recognizing and avoiding predatory OA journals, journals because authors are unaware of what predatory OA there are a number of other approaches that can be used to journals are (although a quantitative understanding of naïvety hinder the progression and spread of predatory OA journals among the aquaculture community is required) and wish to in the field of aquaculture and beyond. For example, Lalu publish their work OA for as cheap a fee as possible. Thus, et al. (2017) suggested that educating researchers on how because predatory OA journals can negatively impact the to identify probable predatory OA journals, and providing reputation of individual researchers (Castillo, 2013; although this incentives for publishing in legitimate journals and disincentives

Frontiers in Marine Science | www.frontiersin.org 9 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

FIGURE 4 | Boxplot representation of demographic article level metrics for indexed (gray boxes) and non-indexed OA (white boxes) journals, including the proportion of articles containing corresponding authors (A) and other authors (i.e., all authors other than the corresponding author) (B) with affiliations in developing countries. Asterisks denote significant differences between journal types; ns denotes non-significant differences (see Table 4). n = 6 per group. for publishing in predatory journals, could help to combat of the gold OA model may not be the most realistic solution and predatory OA publishing. These suggestions could be aided may hinder the progression of and transition to open science. by peer-to-peer and institutional education and outreach, Thus, the approaches mentioned in the previous paragraph may and diligent assessments of researcher publication records by be more effective. institutions and funding agencies; institutional support and While systematic changes will be necessary to eliminate education can also aid graduate students and naïve researchers predatory OA publishing, individual researchers can contribute who wish to ethically publish OA (Beaubien and Eckard, to the cause by being educated about predatory OA journals, 2014). Institutions and funding agencies could also explore and using the “think, check, submit” method when submitting a mandatory archiving in institutional repositories, coupled with manuscript to OA journals (http://thinkchecksubmit.org/). This critical assessments of articles submitted to those repositories method first encourages researchers to think about the journal to monitor and discourage researchers from publishing in they are about to submit their work to and closely consider predatory OA journals (Yessirkepov et al., 2015; Seethapathy whether or not the journal seems legitimate and if their work et al., 2016). Likewise, hiring committees, grant providers, fits in the journal. Next, submitting authors are guided to check international committees, and similar associations need to be the legitimacy of the journal. This can be done using the “think, diligent in assessing researcher credentials to avoid hiring check, submit” checklist (http://thinkchecksubmit.org/check/) and promoting “researchers” that may exploit the lack of and cross-checking the journal or publisher in the DOAJ and rigorous peer review in predatory OA journals to enhance other quality-controlled bibliometric databases. Researchers can their CV. also use the attributes elucidated herein to assess attributes of a An economic shift in the gold OA model could also help candidate OA journal for their work to determine if the journal to alleviate the impact of predatory OA aquaculture journals. may be predatory (see below). Finally, submitting authors are The current state of the gold OA model unfortunately drives encouraged to submit their work only if they answered yes to the exploitation of scientific publishing for monetary gain, most of the questions in the “think, check, submit” checklist. In where an author fee-per-paper model incentivizes higher rates addition, aquaculture researchers can use the criteria herein to of publication and lower rigor. It is important to recognize, distinguish likely predatory journals from legitimate ones. however, that traditional subscription-based publishers also exploit scientific publishing for monetary gain (in most cases, Limitations of the Present Study much more efficiently than OA journals). Given the financial In this study, we found a number of journal- and article- requirements of OA publishing and the difficulty in removing the level attributes that statistically differentiated non-indexed OA associated costs (Hoyt, 2017), removing the financial component aquaculture journals from indexed ones and suggested that

Frontiers in Marine Science | www.frontiersin.org 10 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

TABLE 4 | Results of linear mixed effects models and AIC/log likelihood model comparisons for effects of journal type and journal on demographic article-level metrics.

LME AIC/log likelihood comparison

numDF denDF F-value p-value df AIC BIC logLik L.Ratio p-value

PROPORTION OF CORRESPONDING AUTHORS FROM DEVELOPING COUNTRIES (Intercept) 1 10 76.14 <0.001 w/random 4 15.86 17.08 −3.93 Journaltype 1 10 5.55 0.040 w/orandom 3 13.86 14.77 −3.93 <0.001 >0.999 PROPORTION OF ADDITIONAL AUTHORS FROM DEVELOPING COUNTRIES (Intercept) 1 10 59.75 <0.001 w/random 4 18.65 19.86 −5.32 Journaltype 1 10 2.37 0.155 w/orandom 3 16.65 17.55 −5.32 <0.001 >0.999

Bolded text denotes statistically significant main effects at α ≤ 0.05. non-indexed OA journals may likely be predatory. However, indexation); using the “think, check, submit” method can it is important to recognize that assessing a single attribute be useful in this regard (see section Minimizing the Impact (as highlighted here or in other studies) is insufficient for of Potential Predatory OA Journals in Aquaculture). Future determining whether or not an OA aquaculture journal is studies exploring the variation in article scientific quality within predatory (as not all non-indexed OA journals are predatory; predatory and non-predatory OA journals would provide a more Beaubien and Eckard, 2014); particularly given that the random detailed understanding of the scientific quality of predatory OA factor of journal had a significant effect on most of our journal articles. Furthermore, experimental studies with blinded article-level metrics. Thus, researchers need to carefully examine reviewers assessing the scientific quality of articles from the non- multiple journal attributes in order to conclusively and assuredly indexed OA and indexed journals (compared to an assessment of determine whether or not a given OA journal is likely predatory. aggregate proxies for scientific rigor) would provide more direct Based on our analysis, the aquaculture journals that appear most evidence as to whether or not articles in non-indexed OA journals prominently on Google searches use the gold OA model, were are truly of lesser scientific quality. established after 2010, publish <20 articles per year, and have an While this study documents a variety of quantitative attributes APC under $1,000. In addition, these journals are not included of non-indexed OA aquaculture journals and suggests that they in any of the three quality-controlled bibliometric databases can be predatory, we did not include all potential attributes used herein (DOAJ, Clarivate Analytics, and Scopus). We thus of predatory OA journals. Indeed a number of predatory OA suggest that OA aquaculture journals with this combination journal attributes are evident that we did not incorporate into our of attributes are most likely predatory (as defined previously) analysis, including illustrative aspects of journal covers, unusually and we encourage all authors to assess OA publishing options large editorial boards filled with famous names and/or editors with care. It is critical to stress here, however, that it is the without an online presence, strange/broad subject categories, combination of these attributes that characterize predatory OA and poorly designed websites with many typos and no secure aquaculture journals, and any single attribute cannot be used payment options for open access fees (Hunziker, 2017). As such, to define a journal as predatory. Taking additional steps such future studies objectively assessing the attributes of non-indexed as referring to the DOAJ (https://doaj.org/; and other quality- OA and well-established, indexed journals would benefit from controlled databases such as Web of Science, Clarivate Analytics, incorporating additional attributes such as those highlighted and Scopus) prior to submitting a paper to an OA aquaculture above. journal would also serve researchers well in avoiding predatory Lastly, while this paper focuses on aquaculture research, the OA journals (Günaydin and Dogan, 2015). broader applicability of our results to other fields of research It should also be noted that this study does not suggest remains to be determined. Given our robust design and the that all non-indexed OA aquaculture journals are predatory by similarities between our results and the documented nature of nature, or that all articles contained in predatory OA journals predatory OA journals (Bohannon, 2013), it seems likely that are necessarily of low quality. Thus, our statistical approach our results are applicable to other fields of research. Subtle here is only indicative of potential predatory OA journals. differences in publishing practices, however, could theoretically Nonetheless, it is clear that naïve researchers may submit high render our results invalid for other research disciplines. Future quality work to predatory OA journals. This does not excuse studies applying our approach to objectively assess non-indexed predatory OA journals on the basis that they sometimes publish OA and well-established, indexed journals in other fields of good research, but rather highlights that good science may research are required to determine the breadth of applicability get lost or interpreted as poor science when it is published that our results provide. in such journals. Thus, it is critical for researchers to check attributes of the journals they are submitting to in order to CONCLUSIONS determine whether or not a given OA journal provides an adequate range of scholarly publishing services (e.g., solid peer Our findings suggest that non-scientists are likely to encounter review, participation in archiving, quality-controlled bibliometric predatory OA aquaculture journals in online searches given

Frontiers in Marine Science | www.frontiersin.org 11 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

FIGURE 5 | Diagrammatic of three major ways in which predatory OA publishing threatens science (based on our results), and how such threats apply directly to aquaculture research. that non-indexed OA journals were characteristic of predatory predatory OA journals’ impact on aquaculture research and OA journals and were three times more abundant than the aquaculture industry at large; this can be achieved through indexed journals in the first 10 SERPs of our search. Overall, institutional and peer-to-peer education regarding attributes of non-indexed OA journals were found to use the gold OA model, predatory OA journals (compared to non-predatory journals) were established more recently, and were be less scientifically and promoting the use of databases such as the DOAJ. rigorous than indexed aquaculture journals. Taken together, Additionally, the removal of the pay-per-article aspect of the gold these results suggest that predatory OA publishing may pose OA model would eliminate the purpose of predatory OA journals a threat to aquaculture research and needs to be considered and the negative effects that they inflict on aquaculture. The seriously by the aquaculture research community and industry implementation of such recommendations could help remove as a whole. The awareness and recognition of predatory OA the questionable practices of predatory OA publishers from the aquaculture journals by individual researchers, students, and field of aquaculture research, thus maintaining the integrity of various aquaculture stakeholders is paramount for minimizing the field, moving toward an appropriate public perception of

Frontiers in Marine Science | www.frontiersin.org 12 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals aquaculture in a post-truth era, and ensuring an adequate and and institutional postdoctoral fellowships of RD productive contribution of aquaculture to global food security for and HF. the future. ACKNOWLEDGMENTS AUTHOR CONTRIBUTIONS We are grateful to Dr. Luc A. Comeau and Rémi Sonier at JC: Conceived and helped develop the idea, contributed to the Gulf Fisheries Centre (Fisheries and Oceans Canada) for study design, collected and analyzed data, and wrote and constructive feedback on an earlier draft of this manuscript. revised the manuscript; RD and HF: Contributed to idea We also wish to thank the two reviewers for their substantive development, study design, data collection, and manuscript comments which helped improve the manuscript. revision; RD: Also contributed to data checks and analytical confirmation. SUPPLEMENTARY MATERIAL FUNDING All raw data and R-code are available in the Supplementary Material, which are hosted on GitHub (https://github.com/remi- This project was funded by a NSERC Visiting Postdoctoral daigle/PredPubAquaculture) and archived on Zenodo (https:// Fellowship to JC (Grant number 501540-2016) zenodo.org/record/1188938#.Wp2cXfnwZpg).

REFERENCES Castree, N. (2017). Anthropocene: social science misconstrued. Nature 541:289. doi: 10.1038/541289c Adie, E. (2014). Attention! A study of open access vs. non-open access articles. Christopher, M. M., and Young, K. M. (2015). Awareness of “predatory” open- Figshare. doi: 10.6084/m9.figshare.1213690 access journals among prospective veterinary and medical authors attending Ahrens, W. H., Cox, D. J., and Budhwar, G. (1990). Use of the arcsin and square scientific writing workshops. Front. Vet. Sci. 2:22. doi: 10.3389/fvets.2015. root transformations for subjectively determined percentage data. Weed Sci. 38, 00022 452–458. Clements, J. C. (2017). Open access articles receive more citations in hybrid marine Antelman, K. (2004). Do open-access articles have a greater research impact? Coll. ecology journals. FACETS 2, 1–14. doi: 10.1139/facets-2016-0032 Res. Libr. 65, 372–382. Craig, I. D., Plume, A. M., McVeigh, M. E., Pringle, J., and Amin, M. (2007). Bacher, K. (2015). Perception and Misconceptions of Aquaculture: A Global Do open access articles have greater citation impact? A critical review of the Overview. GLOBEFISH Research Programme, Vol. 120. Rome: Food and literature. J. Informetr. 1, 239–248. doi: 10.1016/j.joi.2007.04.001 Agriculture Organization of the United Nations. d’Ancona, M. (2017). Post-Truth: The War on Truth and How to Fight Back. Balehegn, M. (2017). Increased publication in predatory journals by developing London, UK: Ebury Press. countries’ institutions: what it entails? And what can be done? Int. Inf. Libr. D’Anna, L. M., and Murray, G. D. (2015). Perceptions of shellfish aquaculture in Rev. 49, 97–100. doi: 10.1080/10572317.2016.1278188 British Columbia and implications for well-being in marine social-ecological Beall, J. (2012). Predatory publishers are corrupting open access. Nature 489:179. systems. Ecol. Soc. 20:57. doi: 10.5751/ES-07319-200157 doi: 10.1038/489179a Dockrill, P. (2017). A Neuroscientist Just Tricked 4 Dodgy Journals into Accepting a Beall,. J. (2013). Predatory publishing is just one of the consequences of gold open Fake Paper on ’Midi-Chlorians’. ScienceAlert. Available online at: https://www. access. Learn. Publ. 26, 79–84. doi: 10.1087/20130203 sciencealert.com/a-neuroscientist-just-tricked-4-journals-into-accepting-a- Beall, J. (2016). Dangerous predatory publishers threaten medical research. J. fake-paper-on-midi-chlorians (Accessed July 24, 2017). Korean Med. Sci. 31, 1511–1513. doi: 10.3346/jkms.2016.31.10.1511 Ellis, E., Maslin, M., Boivin, N., and Bauer, A. (2016). Involve social scientists Beaubien, S., and Eckard, M. (2014). Addressing faculty publishing concerns in defining the Anthropocene. Nature 540, 192–193. doi: 10.1038/54 with open access journal quality indicators. J. Libr. Schol. Comm. 2:eP1133. 0192a doi: 10.7710/2162-3309.1133 Eysenbach, G. (2006). Citation advantage of open access articles. PLoS Biol. 4:e157. Benkendorff, K. (2009). “Aquaculture and the production of pharmaceuticals and doi: 10.1371/journal.pbio.0040157 nutraceuticals,” in New Technologies in Aquaculture: Improving Production FAO (2016). The State of the World Fisheries and Aquaculture: Contributing to Food Efficiency, Quality and Environmental Management, eds G. Burnell, and G. Security and Nutrition for All. Rome: Food and Agriculture Organization of the Allan (Oxford: Woodhead Publishing), 866–891. United Nations. Berger, M. (2017). “Everything you ever wanted to know about predatory Fisheries and Oceans Canada (2005). Overview: Qualitative Research Exploring publishing but were afraid to ask,” in Association of College and Research Canadians’ Perceptions, Attitudes and Concerns Towards Aquaculture. Fisheries Libraries. At the Helm: Leading Transformation: The Proceedings of the ACRL and Oceans Canada, Reports and Publications. Available online at: http://www. 2017 Conference, March 22–25, 2017 Baltimore, MD, ed D. M. Mueller dfo-mpo.gc.ca/por-rop/focus-aquaculture-eng.htm (Accessed July 10, 2017). (Chicago, IL: Association of College and Research Libraries), 206–217. Fox, S., and Duggan, M. (2013). Health Online 2013. Washington, DC: Pew Björk, B. C., Laakso, M., Welling, P., and Paetau, P. (2014). Anatomy of green open Research Center’s Internet & American Life Project. access. J. Assoc. Inf. Sci. Technol. 65, 237–250. doi: 10.1002/asi.22963 Froehlich, H. E., Gentry, R. R., Rust, M. B., Grimm, D., and Halpern, B. Bohannon, J. (2013). Who’s afraid of peer review? Science 342, 60–65. S. (2017). Public perceptions of aquaculture: evaluating spatiotemporal doi: 10.1126/science.342.6154.60 patterns of sentiment around the world. PLoS ONE 12:e0169281. Burnham, K. P., and Anderson, D. R. (2002). Model Selection and Multimodel doi: 10.1371/journal.pone.0169281 Inference: A Practical Information-Theoretic Approach, 2nd Edition. New York, Gargouri, Y., Hajjem, C., Larivière, V., Gingras, Y., Carr, L., Brody, T., et al. (2010). NY: Springer-Verlag. Self-selected or mandated, open access increases citation impact for higher Butler, D. (2013). Investigating journals: the dark side of publishing. Nature 495, quality research. PLoS ONE 5:e13636. doi: 10.1371/journal.pone.0013636 433–435. doi: 10.1038/495433a Gasparyan, A. Y., Nurmashev, B., Udovik, E. E., Koroleva, A. M., and Kitas, G. D. Castillo, M. (2013). Predators and cranks. Am. J. Neuroradiol. 34, 2051–2052. (2017). Predatory publishing is a threat to non-mainstream science. J. Korean doi: 10.3174/ajnr.A3774 Med. Sci. 32, 713–717. doi: 10.3346/jkms.2017.32.5.713

Frontiers in Marine Science | www.frontiersin.org 13 March 2018 | Volume 5 | Article 106 Clements et al. Evaluating Open Access Aquaculture Journals

Günaydin, G. P., and Dogan, N. Ö. (2015). A growing threat for quality of Indian science and education. Curr. Sci. 111, 1759–1764. academicians: fake and predatory journals. J. Acad. Emerg. Med. 14, 94–96. doi: 10.18520/cs/v111/i11/1759-1764 doi: 10.5152/jaem.2015.48569 Shao, Z. J. (2001). Aquaculture pharmaceuticals and biologicals: current Gutierrez, F. R., Beall, J., and Forero, D. A. (2015). Spurious alternative impact perspectives and future possibilities. Adv. Drug Deliv. Rev. 50, 229–243. factors: the scale of the problem from an academic perspective. Bioessays 37, doi: 10.1016/S0169-409X(01)00159-4 474–476. doi: 10.1002/bies.201500011 Shen, C., and Björk, B. C. (2015). “Predatory” open access: a longitudinal Hajjem, C., Harnad, S., and Gingras, Y. (2005). Ten-year cross disciplinary study of article volumes and market characteristics. BMC Med. 13:230. comparison of the growth of open access and how it increases research citation doi: 10.1186/s12916-015-0469-2 impact. Bull. IEEE Comp. Soc. Tech. Comm. Data Eng. 28, 39–47. Sorokowski, P., Kulczycki, E., Sorokowska, A., and Pisanski, K. (2017). Predatory Harris, E. (2004). Building scientific capacity in developing nations. EMBO Rep. 5, journals recruit fake editor. Nature 543, 481–483. doi: 10.1038/543481a 7–11. doi: 10.1038/sj.embor.7400058 Swan, A., Gargouri, Y., Hunt, M., and Harnad, S. (2015). Open access policy: Holmes, G., Barber, J., and Lundershausen, J. (2017). Anthropocene: be wary of numbers, analysis, effectiveness. arXiv:1504.02261. social impact. Nature 541:464. doi: 10.1038/541464b Swauger, S. (2017). Open access, power, and privilege: a response to “What Hoyt, J. (2017). Untitled. Twitter. Available online at: https://twitter.com/ I learned from predatory publishing”. Coll. Res. Libr. News 78, 603–606. jasonHoyt/status/879624241817296896 (Accessed July 14, 2017). doi: 10.5860/crln.78.11.603 Hunziker, R. (2017). Avoiding predatory publishers in the post-Beall world: tips for United Nations (2017). World Economic Situation and Prospects 2017. New York, writers and editors. AMWA J. 32, 113–115. NY: United Nations. Jalalian, M., and Mahboobi, H. (2014). Hijacked journals and predatory Van Nuland, S. E., and Rogers, K. A. (2016). Academic nightmares: predatory publishers: is there a need to re-think how to assess the quality of publishing. Anat. Sci. Educ. 10, 392–394. doi: 10.1002/ase.1671 academic research? Walailak J. Sci. Technol. 11, 389–394. doi: 10.14456/wjst. Vinny, P. W., Vishnu, V. Y., and Lal, V. (2016). Trends in scientific publishing: dark 2014.16 clouds loom large. J. Neurol. Sci. 363, 119–120. doi: 10.1016/j.jns.2016.02.040 Keyes, R. (2004). The Post-Truth Era: Dishonesty and Deception in Contemporary Wakefield, A. J., Murch, S. H., Anthony, A., Linnell, J., Casson, D. M., Malik, Life. New York, NY: St. Martin’s Press. M., et al. (1998). Ileal-lymphoid -nodular hyperplasia, non-specific colitis, Lalu, M. M., Shamseer, L., Cobey, K. D., and Moher, D. (2017). How stakeholders and pervasive developmental disorder in children. Lancet 351, 637–641. can respond to the rise of predatory journals. Nat. Hum. Behav. 1, 852–855. doi: 10.1016/S0140-6736(97)11096-0 doi: 10.1038/s41562-017-0257-4 Wang, X., Liu, C., Mao, W., and Fang, Z. (2015). The open access advantage Lönnestedt, O. M., and Eklöv, P. (2016). Environmentally relevant concentrations considering citation, article usage and social media attention. Scientometrics of microplastic particles influence larval fish ecology. Science 352, 1213–1216. 103, 555–564. doi: 10.1007/s11192-015-1547-0 doi: 10.1126/science.aad8828 Ware, M., and Mabe, M. (2015). The STM Report: An Overview of Scientific McKiernan, E. C., Bourne, P. E., Brown, C. T., Buck, S., Kenall, A., Lin, J., and Scholarly Journal Publishing, Fourth Edition. The Hague: International et al. (2016). How open science helps researchers succeed. eLife 5:e16800. Association of Scientific, Technical and Medical Publishers. doi: 10.7554/eLife.16800 Warton, D. I., and Hui, F. K. C. (2011). The arcsin is asinine: the analysis of McNutt, M. (2016). My love hate of Sci Hub. Science 352:497. proportions in ecology. Ecology 92, 3–10. doi: 10.1890/10-0340.1 doi: 10.1126/science.aaf9419 Xia, J., Harmon, J. L., Connolly, K. G., Donnelly, R. M., Anderson, M. R., and Moher, S., and Srivastava, A. (2015). You are invited to submit. BMC Med. 13:180. Howard, H. A. (2015). Who publishes in “predatory” journals? J. Assoc. Inf. doi: 10.1186/s12916-015-0423-3 Sci. Technol. 66, 1406–1417. doi: 10.1002/asi.23265 Pauly, D., and Zeller, D. (2016). Catch reconstructions reveal that global marine Yessirkepov, M., Nurmashev, B., and Anartayeva, M. (2015). A scopus-based fisheries catches are higher than reported and declining. Nat. Commun. analysis of publication activity in Kazakhstan from 2010 to 2015: positive 7:10244. doi: 10.1038/ncomms10244 trends, concerns, and possible solutions. J. Korean Med. Sci. 30, 1915–1919. Pinheiro, J., Bates, D., DebRoy, S., and Sarkar, D. (2017). nlme: Linear and doi: 10.3346/jkms.2015.30.12.1915 Nonlinear Mixed Effects Models. Available online at: https://CRAN.R-project. org/package=nlme (Accessed July 11, 2017). Conflict of Interest Statement: The authors declare that the research was Purcell, K., Brenner, J., and Rainie, L. (2012). Search Engine Use 2012. conducted in the absence of any commercial or financial relationships that could Washington, DC: Pew Research Center’s Internet & American Life be construed as a potential conflict of interest. Project. R Development Core Team (2017) R: A Language and Environment for Statistical Copyright © 2018 Clements, Daigle and Froehlich. This is an open-access article Computing, Version 3.4.1. Vienna: R Foundation for Statistical Computing. distributed under the terms of the Creative Commons Attribution License (CC Richter, F. (2013). 1.17 Billion People Use Google Search. Available online at: https:// BY). The use, distribution or reproduction in other forums is permitted, provided www.statista.com/chart/899/unique-users-of-search-engines-in-december- the original author(s) and the copyright owner are credited and that the original 2012 (Accessed September 28, 2017). publication in this journal is cited, in accordance with accepted academic practice. Seethapathy, G. S., Santhosh Kumar, J. U., and Hareesha, A. (2016). No use, distribution or reproduction is permitted which does not comply with these India’s scientific publication in predatory journals: need for regulating terms.

Frontiers in Marine Science | www.frontiersin.org 14 March 2018 | Volume 5 | Article 106